COMPARISON OF PERFORMANCE OF RBF AND MLP NEURAL NETWORKS FOR RESULTS OF SIMULTANEOUS HEAT & MASS TRANSFER

In most of the chemical engineering processes, the phenomena of mass and heat transfer are among their inseparable parts. In the present paper, simultaneous heat and mass transfer has been studied experimentally by a laboratory setup. In this apparatus, the existence of condensation and evaporation created due to heat transfer causes mass transfer and finally influences the coefficient of heat transfer. Also, mass transfer changes heat distribution in heat transfer phenomena that causes total change in heat flux. Thus, in this apparatus, some experiments have been carried out through changing different parameters such as temperature and flow rate for two fluids, water and air. Also in this paper, it is tried to compare the function of these systems with each other and the results acquired from the experiments as well as the capacity of these systems in analyzing the results were studied through artificial neural networks. From among the neural networks used in this paper, we may refer to RBF, MLP networks. The studies indicate that the MLP network is not able to predict properly due to lack of any facility for noise filtration and RBF network has the best function due to having a stronger theoretical basis.